INTRODUCTION
Resection margin (RM) involvement after curative intent resection for gastric cancer results in a poor prognosis and deprives the patient of the chance for cure. Although intraoperative frozen biopsy is a reliable and widely used method, the false-negative rate of that is 0.5% to 4.7% [1,2]. However, there are few cases of reoperation to achieve a negative RM because it should guarantee tolerable morbidity and achievement of negative margins [3]. Recently, we performed remnant total gastrectomy and supra-ampullary duodenectomy in a patient with positive proximal and distal RMs after distal gastrectomy for gastric cancer. Herein, we describe the perioperative procedure and management. Informed consent was obtained from the patient for being included in the study.
CASE REPORT
A 56-year-old woman with gastric cancer and no other relevant medical history visited the outpatient clinic. She had undergone three cycles of neoadjuvant chemotherapy with FLOT (docetaxel, oxaliplatin, fluorouracil, and leucovorin) regimen in another hospital 1 month ago.
Esophagogastroduodenoscopy before chemotherapy had revealed an ulcerative lesion with signet ring cell carcinoma in the lower body of the stomach. Abdominal computed tomography (CT) and positron emission tomography-CT before chemotherapy had shown gastric wall thickening involving the lower body and antrum, with fluorodeoxyglucose uptake. Preoperative positron emission tomography-CT revealed that the previously noted hypermetabolic lesion in the gastric wall was nearly normalized. There was no evidence of lymph node or distant metastases (Fig. 1).
Laparoscopic distal gastrectomy with D2 lymph node dissection and extracorporeal Billroth II reconstruction was performed. No enlarged lymph nodes or peritoneal seeding was observed. Proximal transection was performed 1 cm below the esophagogastric junction on the lesser curvature side, and distal transection was performed 2 cm distal to the pylorus. The gross proximal and distal margins were of 3.5 cm and 2 cm, respectively. Frozen biopsy revealed a negative RM at both ends (Fig. 2). The final pathological diagnosis was ypT3N2M0 with the negative result of washing cytology. The specimen was of Borrmann type IV cancer involving the whole stomach. Some single cancer cells were observed on immunohistochemical staining for cytokeratin in the proximal and distal RMs (Fig. 3).
Two weeks postoperatively, open remnant total gastrectomy and supra-ampullary duodenectomy were performed. Before reoperation, percutaneous transhepatic gallbladder drainage (PTGBD) and angiocatheter placement outside the ampulla of Vater (AoV) via the cystic duct were performed to avoid pancreaticoduodenectomy (PD) and obtain the maximal distal margin (Fig. 4). The operative findings revealed moderate adhesion at the superior pancreatic border and gastrojejunostomy site. Duodenal transection was performed 1 cm above the AoV through palpation of the angiocatheter, following kocherization and detachment of the duodenum from the pancreas. Invagination of the duodenal stump followed the duodenal transection. The resected duodenum was 4 cm long (Fig. 5). Two Jackson-Pratt (JP) drains were inserted around the duodenal stump and behind the esophagojejunostomy.
The final pathology revealed a negative RM at both ends, and there was no carcinoma in the resected duodenum. On a postoperative day 4, upper gastrointestinal series and fistulography by injecting contrast material through the angiocatheter revealed no anastomotic leakage. On postoperative day 8, abdominal CT was performed and no significant fluid collection was found. Accordingly, the JP drain placed behind the esophagojejunostomy and the angiocatheter were removed. On postoperative day 9, the JP drain around the duodenal stump was removed, and the patient was discharged without complications. The PTGBD was removed after clamping for 4 days. The patient received three cycles of adjuvant chemotherapy with the FLOT regimen 6 weeks after the initial operation. She was followed up every 3 months with carcinoembryonic antigen, carbohydrate antigen 19-9, and chest and abdominal CT. Recurrence with peritoneal seeding was detected 12 months later, and the patient has been receiving palliative chemotherapy.
DISCUSSION
On analyzing RMs following intraoperative frozen biopsy for gastric adenocarcinoma, the risk factors for false-negative results were found to be neoadjuvant chemotherapy, signet ring cell carcinoma, diffuse type, and Borrmann type IV carcinoma [1,2].
In our case, all of these risk factors were present. In initial esophagogastroduodenoscopy, there were no specific findings other than ulcerative lesion and fold thickening around the lesion in lower body greater curvature because the gross type was Borrmann type IV, in which cancer cells spread along the submucosa and muscle layer. This was supported by pathologic findings. We focused on the proximal RM intraoperatively because Borrmann type IV cancer often involves the entire stomach. Thus, we transected the stomach extracorporeally for palpation of wall thickening. Although it seemed that RM was less than 5 cm recommended in the guideline, we deferred the decision of RM to frozen section because there is no evidence about RM except utilization of frozen section [4]. However, false-negative results were observed at both ends.
In most studies, the number of R1 cases is 2% to 8%, and reoperation was performed in few of them [3]. For patients with microscopically positive RMs, there is some debate about the survival benefit of reoperation to achieve R0 status in advanced stage although most authors recommended reoperation in the early stage [5]. Unfortunately, our patient had a ypT3N2M0 tumor, and recurrence was detected 12 months postoperatively.
However, it is clear that reoperation is the only treatment option to achieve R0 status with a chance of cure. Therefore, clinicians should explore perioperative management, including optimal operation methods, to prevent severe complications and achieve R0 resection, especially in medically fit patients.
With respect to gastric cancer with esophageal invasion, the abdominal-transhiatal approach is feasible if the length of the esophageal resection is within 6 cm. No survival benefit of extended aggressive surgery was observed for esophageal involvement of ≥3 cm in subcardia cancer [6].
In terms of gastric cancer with duodenal invasion, Kakeji et al. [7] suggested that the first portion of the duodenum is 4 to 5 cm in length, which is sufficient to resect 3 to 4 cm of the duodenum beyond the pylorus, and PD can be considered when the lesion extends >3 cm beyond the pyloric ring. However, the postoperative morbidity rate of PD was 23% to 74%, the pancreatic leak rate was 24.5%, and the mortality rate was 1.4% in the pooled analysis [8]. Yonemura et al. [9] reported no survival benefit of PD for tumors with duodenal invasion. On the other hand, Ajisaka et al. [10] concluded that curative resection including PD improved the prognosis of patients with duodenal invasion, except in nodal type duodenal invasion, which involves nodal metastatic lesions around the pancreatic head and is the longest duodenal invasion type with a mean length of 3.6 cm [10].
In our case, CT before reoperation revealed that the length from the duodenal stump to the AoV was approximately 3 cm, and accordingly, we planned further duodenal resection, not PD. However, it is difficult to guarantee no injury of AoV during detachment of the duodenum from the pancreas for maximal achievement of duodenal margin beyond 4 cm. This confidence is even more necessary in reoperation.
Recently, Di Saverio et al. [11] introduced pancreas-sparing, ampulla-preserving D1–D2 duodenectomy for emergency treatment of major duodenal perforations. They performed cholecystectomy and tube placement from the cystic duct to the duodenal lumen outside the AoV for proximal duodenectomy with preservation of the AoV.
We transformed this surgical technique to a preoperative intervention to avoid pancreatico-duodenectomy and obtain the maximal distal margin. A drainage catheter was placed in the duodenum outside the AoV via the cystic duct after PTGBD was attempted. However, the cystic duct was too thin for the drainage catheter to pass through in this patient. Finally, PTGBD and angiocatheter placement were performed in the duodenum outside the AoV via the cystic duct. These were useful not only for duodenectomy but also for postoperative management. Although gallstone formation is one of the common postoperative complications of gastric cancer surgery, cholecystectomy was not performed in this patient because the role of prophylactic cholecystectomy remains controversial [12,13]. The total length of the resected duodenum was 6 cm, which was sufficient for the free margin of tumors with duodenal invasion.
In summary, supra-ampullary duodenectomy can be considered the most suitable treatment in terms of a chance for cure in patients with positive distal margin after subtotal gastrectomy for gastric cancer. Preoperative percutaneous catheter placement in the duodenum, outside the AoV, via the cystic duct, can aid re-excision to obtain sufficient margins and relieve the burden on the operator in terms of postoperative management.